1. Gene Expression Matries
|
Level
|
Quantifier
|
Metrics
|
filePath
|
|
Transcript
|
RSEM
|
Raw counts
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_transcriptLevel.RSEM_Count.txt
|
|
Transcript
|
RSEM
|
TPM
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_transcriptLevel.RSEM_TPM.txt
|
|
Transcript
|
RSEM
|
FPKM
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_transcriptLevel.RSEM_FPKM.txt
|
|
Transcript
|
Salmon
|
Raw counts
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_transcriptLevel.Salmon_Count.txt
|
|
Transcript
|
Salmon
|
TPM
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_transcriptLevel.Salmon_TPM.txt
|
|
Gene
|
RSEM
|
Raw counts
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_geneLevel.RSEM_Count.txt
|
|
Gene
|
RSEM
|
TPM
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_geneLevel.RSEM_TPM.txt
|
|
Gene
|
RSEM
|
FPKM
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_geneLevel.RSEM_FPKM.txt
|
|
Gene
|
Salmon
|
Raw counts
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_geneLevel.Salmon_Count.txt
|
|
Gene
|
Salmon
|
TPM
|
/research_jude/rgs01_jude/groups/yu3grp/projects/software_JY/yu3grp/conda_env/bulkRNAseq_2025/pipeline/testdata/Quantification/_research_jude_rgs01_jude_groups_yu3grp_projects_software_JY_yu3grp_conda_env_bulkRNAseq_2025_pipeline_databases_hg38_gencode.release48/01_expressMatrix_geneLevel.Salmon_TPM.txt
|
2. Alignment statistics
Alignment statistics helps to tell: 1）if
you have sequenced enough reads? Usually, >50M
mapped reads and >25M uniquely-mapped reads
are expected; 2）if your reads have been over-trimmed?
Usually, >90% Percentage of Filtered Reads is
expected. Over-trimming could happen when the insert size is too small
or you picked the wrong Phred score encoding method; 3）if your
samples are contaminated? Usually, >75% Percentage
of Mapped Reads is expected. The most common reason for low
mapping rate is the contamination of either DNA or rRNA.
|
Sample Name
|
Number of Total Reads
|
Number of Filtered Reads
|
Percentage of Filtered Reads
|
Number of Mapped Reads
|
Percentage of Mapped Reads
|
Number of Uniquely-mapped Reads
|
Percentage of Uniquely-mapped Reads
|
|
sample1
|
38820746
|
38327088
|
98.73%
|
31672982
|
82.64%
|
29326030
|
92.59%
|
|
sample5
|
27712150
|
27534450
|
99.36%
|
23095284
|
83.88%
|
21399082
|
92.66%
|
|
sample2
|
16005450
|
16004460
|
99.99%
|
1125972
|
7.04%
|
800744
|
71.12%
|
* Percentage of Uniquely-mapped Reads = Number of
Uniquely-mapped Reads / Number of Mapped Reads
|
* Percentage of Mapped Reads = Number of Mapped Reads /
Number of Filtered Reads
|
3. Quantification statistics
Quantification statistics helps to tell how accurate
the quantification results are: 1）how many transcripts/genes
were identified confidently? Usually, >65K
transcripts and/or >15K genes are expected;
2）how accurate the quantification is? Usually,
>0.85 correlation coefficient is expected.
3.1 Transcript level
|
Sample Name
|
Identified by RSEM
|
Identified by Salmon
|
Identified by Both
|
Coef_Pearson
|
Pval_Pearson
|
rho_Spearman
|
Pval_Spearman
|
|
sample5
|
85686
|
80064
|
73906
|
0.985
|
0
|
0.9712
|
0
|
|
sample1
|
91688
|
87064
|
80009
|
0.979
|
0
|
0.9698
|
0
|
|
sample2
|
17121
|
22611
|
11230
|
0.2311
|
4.78723092812193e-136
|
0.8404
|
0
|
* Only co-identified transcripts/genes were used in
correlation analysis.
|
3.2 Gene level
|
Sample Name
|
Identified by RSEM
|
Identified by Salmon
|
Identified by Both
|
Coef_Pearson
|
Pval_Pearson
|
rho_Spearman
|
Pval_Spearman
|
|
sample5
|
21237
|
21422
|
20535
|
0.9856
|
0
|
0.9883
|
0
|
|
sample1
|
22114
|
22431
|
21364
|
0.9798
|
0
|
0.9877
|
0
|
|
sample2
|
9735
|
12847
|
9233
|
0.2505
|
4.22481223109004e-132
|
0.8377
|
0
|
* Only co-identified transcripts/genes were used in
correlation analysis.
|
4. Biotype distribution
Biotype distribution provides the composition of
types of identified transcripts and genes. Since different bioytpes of
transcripts/genes vary hugely in length, GC-content and other
properties, Biotype distribution could serve as a
measure of quantification quality. Empirically, 1) for
total RNA libraries, the protein_coding transcripts and
genes should account for >40% and >
50%, respectively; 2) for mRNA
libraries, the protein_coding transcripts and genes
should accounts for >80. For more details about the
biotypes: https://www.gencodegenes.org/pages/biotypes.html.
4.1 Transcript level
|
sampleName
|
protein_coding
|
retained_intron
|
lncRNA
|
protein_coding_CDS_not_defined
|
nonsense_mediated_decay
|
Others
|
processed_pseudogene
|
|
sample2
|
55.57%
|
14.65%
|
2.46%
|
12.31%
|
10.17%
|
1.35%
|
3.49%
|
|
sample5
|
44.29%
|
20.2%
|
15.88%
|
9.17%
|
8.6%
|
1.87%
|
NA
|
|
sample1
|
43.19%
|
20.26%
|
16.79%
|
9.15%
|
8.71%
|
1.9%
|
NA
|
* The biotypes of <1% were marked as ‘NA’ and merged into
‘Others’.
|
4.2 Gene level
|
sampleName
|
protein_coding
|
lncRNA
|
processed_pseudogene
|
TEC
|
Others
|
|
sample2
|
87.68%
|
3.91%
|
6.14%
|
NA
|
2.26%
|
|
sample5
|
64.97%
|
27.92%
|
3.51%
|
1.17%
|
2.43%
|
|
sample1
|
63.41%
|
29.16%
|
3.57%
|
1.23%
|
2.63%
|
* The biotypes of <1% were marked as ‘NA’ and merged into
‘Others’.
|
5. Genebody coverage statistics
Genebody coverage statistics calculates the RNA-Seq
reads overage over gene body. 1) Mean of Coverage is
the average coverage of all 100 bins of gene body. Usually,
>0.7 is expected. 2) Coefficient of
Skewness is a measure of the asymmetry of the distribution of
gene body coverage. Fisher’s moment coefficient of
skewness was calculated by default. The closer to 0, the more
symmetric. For more details: https://en.wikipedia.org/wiki/Skewness.
|
sampleName
|
meanCoverage
|
coefSkewness
|
|
sample5
|
0.8555260
|
0.0709431
|
|
sample1
|
0.8449504
|
0.0756538
|
|
sample2
|
0.5527551
|
0.3571358
|